So, why did you move from - or is this just securing your name on .café??
are you going to talk about different things on this instance?

did I miss something? (Have been out teaching tech for a couple of days...)

@eileenb The primary incentive was that @TheAdmin has apparently subbed in a local CSS stylesheet which results in all content being half-width on my tablet in portrait mode. That's how I view things when I'm composing (or reading-and-composing) things, which is far less than ideal. I've been requesting he remove that for a couple of days now, but our views differ.

I'd been meaning to try another instance in any regard, for a few reasons, out of curiosity.

I'd joined MC on the premise that there would be more activity visible on a larger instance, and it was the largest instance accepting registrations when I joined Mastodon. That prediction appears generally true.

But more != better, and with the tremendous amount of Japanese participation in Mastodon, neither the Local nor Federated timelines are of much interest or use to me. is far smaller (~50 users, 640 max) and has an academic/intellectual bent.


@dredmorbius aah - that makes sense - I think all the Japanese users flocked there after a media article - trouble is, you can't tell what mastodon is like until you are in an instance. its still the wild west here with no form of education for folks joining. perhaps all admins should describe the activity on their instance more fully :)

@eileenb Very good points, yes. I've been suggesting protocols for describing and sharing information on instances, standardising Admin contact points (and figuring out how not to turn those into the spampits that standard postmaster & webmaster addresses are in email), etc.

I'd also like to see a blocklist / whitelist exchange protocol so there's transparency between instances on what is or isn't embargoed. CoC as well, likely.

@eileenb There was a /very/ good article in "ASCII" magazine, by a long-time computer researcher in Japan.

It was apparently ... persuasive.
(Google translate does an excellent job of rendering this.)

I wanted to see what that was like.

1. There's /very/ little local activity. I'm actually most of it since joining, though nobody's complained ... yet.
2. I've apparently encouraged a few other people to join among my acquaintances. About 4-5, so not large numbers absolutely but about 10% of the local instance.
3. The local timeline is all European languages, which I can generally puzzle out.
4. The Federated timeline is /vastly/ less busy.


@dredmorbius you're persuading me...!!
reluctant to boost your post in case your new instance gets flooded too :)

@eileenb Well, the instance is capped at 640 members, for now. I'll suggest to @skk that they might want to consider making joins request or approval-based if a flood starts.

That said: my view is that Mastodon works better if good content /is/ emphasised. Or at least I hope I'm right about that. There's a bit of a discussion at G+ right now over why it's dying and what it might have been. @gideonro wrote a post there that's gone nuts. 350+ reshares.

@dredmorbius @skk @gideonro will read it later thanks. First, weekend stuff like market shopping, baking bread and making cheese ;)

@eileenb Never underprioritise cheesemaking!

(I'll need to ask you about doing that.)

Wups, I exaggerated: 339 reshares. That's ... a /lot/ on G+.

If you look at the Baez re-share, I gave a long rundown on what I've been looking for, and largely failing to find online. Some exceptions, but they're rare.

@gideonro @eileenb

@dredmorbius @eileenb That's the plan for now. Since this is more of a side project (as in: also got another job that pays the bills, not as in: I'm only slightly interested), I figured that a smallish community might be easier on the server and on my time to begin with.

But afterwards, all options remain open: invitation-only instance, opening to other mods/admins, etc.

Still, early days and all that...

@skk @eileenb Understood. I think you're right with the smaller-community idea.

I also think that being /selective/ in that smaller community will pay off as well.

I've had good results with a small (40-50 people) community at G+, of selected people and a set of general participation guidelines.

Similar to these:

@dredmorbius I'm sure there's a way to mute all posts where charset=kanji. even a non programmer like me can suss that out. Twitter does that at the back end...

@eileenb There's no language / charset muting yet, though it is almost certainly requested...

(I also just realised that has its own fork of mastodon.)

@dredmorbius @eileenb

Admittedly, 'curiosity' of other instances was cause enough to look elsewhere. A couple of ideas came from this:

a. I wonder if this allows a subreddit model of sorts, where diversity of "interest-based instances" is followed. Subreddits without Reddit?, and

b. It seems 'natural' to define one's own boundaries. So the ability to associate is just as important as the ability to disassociate, no?

@prole @eileenb The Mastodon model, as it is, allows only a single user-interest mapping. This could change.

The Reddit model allows /multiple/ user-interest mappings. That's inherently more flexible. Though it tends to lose most levels of inter-user bonding, most especially in large (million-plus) subreddits. You'll get some of this in smaller ones, possibly.

I'd really like to see some sense of user + locality + interest overlay + outward and inward content filtering.

@eileenb @prole That would rely heavily on some standard set of channels (and/or sub-channels), with specific content targeted at such. Also an option for, say, recipients-only or local-only distribution: messages which remain /only/ with a given user or instance.

For channels, one or more might be applied to a message, and initial distribution scoped by the interest applied to that channel, say, as a probability 0 < p < 1.

The lower the probability, the larger the overlay.

@eileenb @prole So, say, "sex" or "kittens". Interest is high, so p ~= 1, so the overlay is "local".

For "interior odan paleogeochemistry", p ~= 0, so the overlay is, roughly, "global". That is, it's a low-interest topic, and you probably want to reach just about everyone who's interested in it.

For high-p discussions, the /best/ set of posts / threads (top n for some time period t) are promoted to the next level. The dross stays behind.

@prole @eileenb For moderate p values, say, 0.1 < p < 0.25, you've got the situation where an initial discussion includes /multiple/ locales, but only a subset of total instances.

This scheme admittedly works better where instance sizing is roughly equivalent. I'm not sure how that balances out all told, and need to think about the 3-4 OOM variance extant (1 - 100k profiles/instance). Weighting factors, probably.

Classifications: Library of Congress (reasons prev. stated).

Since what you see on the Federated timeline is /not/ everything posted anywhere on Mastodon (or extended networks), but (generally) what's postsed by profiles /followed by your local instance/, there's a much more effective interest filter from Again, I am contributing a lot to that (something I now have to consider), but I'm appreciating a generally more-sober stream. And more readable one.

This plays into information overload, a growing interest.


My formulation is "rivality is the counterpoint of virality". That is: whilst it's true that any one post or comment can go viral, /only one item/ can be the most-seen (or listened to, or watched, or read) of some time period and audience.

As your audience (and contributor) pool increases, the number of top-10 slots remains constant.

And there's a limit to media consumption per day, either individually or for a population. Somewhere between 10 - 1k items.


Not original, of course. Herbert Simon and Alvin Toffler were writing about this in the 1970s. Peter Gleick's got "when information is abundant, attention is scarce". Toffler goes into depth of psychological implications in "Future Shock".

I think the effects and implications are pretty profound, and have been hacking at my information and attention space for a while (not always consciously or intentionally), with interesting effects.

So it's an experiment.


Sign in to participate in the conversation

Everyone is welcome as long as you follow our code of conduct! Thank you. is maintained by Sujitech, LLC.